The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the user to install the MCP server using npx -y @nataliapc/mcp-openmsx. This involves downloading and executing a remote package from the NPM registry. The package is managed by the same author as the skill.
[DYNAMIC_EXECUTION]: The skill references an advanced tool called openmsx_tcl_cmd which allows for unrestricted execution of Tcl commands within the openMSX environment.
The documentation states that this tool can modify emulator state, access host files, and load scripts.
The skill implements a security control requiring the user to explicitly set an environment variable (OPENMSX_ENABLE_RAW_TCL=true) before the tool is registered.
Instructions warn the agent never to enable this feature itself, emphasizing user control.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted data from the emulator and external sources.
Ingestion points: Data enters the agent's context through screen text capture (screenGetFullText), resource retrieval (read_mcp_resource), and vector database queries (vector_db_query).
Boundary markers: There are no specific delimiters or instruction-bypass warnings defined for data retrieved from the emulator or documentation database.
Capability inventory: The skill can perform memory writes (writeByte, writeBlock), manipulate CPU registers (setRegister), and execute Tcl commands (if enabled).
Sanitization: No explicit sanitization or validation logic is mentioned for data returned by the search tools or screen capture functionality.
[COMMAND_EXECUTION]: The documentation provides multiple shell command examples for configuring various MCP clients (VS Code, Codex) and starting the server in different transport modes. These are standard configuration instructions for this type of integration.

mcp-openmsx-usage